BT’s Huawei Cabinets Give Better VDSL2 “Fibre Broadband” Speed than ECI

Anybody familiar with Openreach’s (BT) roll-out of VDSL2 based ‘up to’ 80Mbps Fibre-to-the-Cabinet (FTTC) tech knows that the operator tends to deploy two different types of Street Cabinet hardware, one from Huawei and another from ECI. But a new study suggests that Huawei’s kit is faster.

The use of two different types of cabinet / hardware is to be expected and simply reflects natural diversity in supply (i.e. don’t put all your eggs in one basket), but at the same time we’ve long noticed that ECI hardware on Openreach’s network has tended to run into trouble, while Huawei’s kit seems to suffer fewer hiccups.

For example, ECI kit didn’t have the best time when Openreach was testing Vectoring to remove cross-talk interference from VDSL2 lines (here) and the recent attempt to roll-out of Physical Retransmission ReTX (G.INP) also had plenty of hassle from the ECI side of things (here).

Similarly we understand that the on-going trial of a 3dB Noise Margin (SNR) may need a bit longer before it’s happy to be deployed with ECI kit (here), but the official deployment isn’t due to begin until March 2017 and so there’s still time to sort that out. All of the above features can aid performance and the fact that ECI tends to struggle certainly won’t help.

Now Thinkbroadband have decided to take a closer look at the performance differences between ECI and Huawei cabinets on VDSL2 / FTTC lines, which finds that a difference does appear to exist. You can see the full results on their website, but here’s a short extract.

Overall the benefit appears to be fairly small, reflecting an extra 1-2Megabits per second for people connected to a Huawei cabinet versus an ECI one and we’re told that it’s a similar pattern for upload speeds.

However this information should be taken with a pinch of salt as every area will be different and in some locations ECI kit may do better. Openreach also appear to be deploying more Huawei cabinets now than ECI (perhaps because of past problems), which may also be distributed differently. Never the less it’s a fascinating little study.